1. Introduction

This chapter provides an introduction to the benefits and features of Mentor, and a summary of each chapter in this manual. It also provides an overview of the process of producing tables from data. For more information go to 1.3 THE BASIC STEPS OF THE REPORT/TABBING PROCESS.

If you need a definition of any terms or concepts, a glossary of terms is included in the Utilities manual. Users of Survent, CfMC’s interviewing package, will be familiar with many of the concepts and terms used in this manual.

1.1 WHAT IS MENTOR?

Mentor is a reporting and cross-tabulation program designed for companies in market analysis, telemarketing, market research and related fields. In addition, Mentor comes with various supporting utilities.

1.2 BENEFITS OF USING MENTOR

There are many advantages to using Mentor, including:

  • Complex feature set
  • Powerful command language
  • Advanced database capabilities
  • Integration with Survent
  • Ease of Use

Each of these benefits is explained in the following sections.

COMPLEX FEATURE SET

Mentor’s primary function is to create tables, but it also does the following:

  • clean data
  • create variables
  • create a database, which includes:
    • variables
    • procedures
    • table elements
    • tables
  • specify bases and filters
  • weight the data in tables
  • use IF-THEN-ELSE logic
  • run statistics, such as:
    • Mean
    • Standard Deviation
    • Standard Error
    • Newman-Keuls
    • T-tests
    • Chi-squares
    • All-possible-pairs significance testing
  • list open-ends
  • show frequency distributions
  • input and output multiple data types, such as:
    • CfMC System files
    • ASCII
    • Binary
    • swapped-binary
    • hex
    • sort data files
  • concatenate data files
  • merge data files
  • create subsets of data files
  • print tables in various formats

This is only a partial list of features, and new features are being added with each release. See the section 1.6 GETTING TECHNICAL SUPPORT at the end of this chapter to see how to get the latest information about Mentor and how to submit a request to the CfMC programming staff.

POWERFUL COMMAND LANGUAGE

Mentor is a full-featured, cross-tabulation and reporting system. It is a program that gives users direct command-line capabilities with which to process research information. This complex program is efficient for the highly skilled spec writer who wishes to use the spec language directly. (New users may want to generate their tables using SCAN, a utility program described in the Utilities manual.)

While simple, straightforward jobs are easy to produce, Mentor also allows you to handle large, complicated jobs as well. You can create boiler-plate specification files and then tailor them for specific jobs. You can also create libraries of questions or specifications for specific clients or jobs.

One powerful feature of the Mentor command language is the ability to join data locations in a variety of ways. Data locations can be combined using a variety of joiners (or connectors) including arithmetic joiners (add, subtract, multiply, divide), vector joiners (WITH, BY), logical joiners (AND, OR, NOT), relational joiners (>, <, =, and < >), and combinations of these. Joining variables allows you to create complex expressions that can be used for table banners, for example. This combining of simple variables into more complex expressions is efficient because the building blocks, once defined as variables, can be reused repeatedly, simply by referring to them by variable name, once they are in the database file.

You can create new categories or manipulate the data using Survent screens or at the spec line. You can combine existing variables to create complex combinations of categories, and give these definitions their own variable names as well. Your definitions can include bases, weights, statistics and significance testing. You also have several different formatting and printing options. You can print a table to your screen, revise it as needed, and then print it again, saving it this time to a print file or sending it to a printer.

The utilities are user-friendly menus that prompt a beginning user through building a simple cross-tab or defining a questionnaire. You can use Survent and Mentor to create your own utilities or interfaces that are tailored to the specific needs of your own shop or clients.

ADVANCED DATABASE CAPABILITIES

Mentor creates and uses a database contained in a DB file. The DB file acts as a library, maintaining definitions of all the various elements processed throughout a complete job. The Mentor database significantly reduces processing time and makes the administration of large, repetitive jobs (such as tracking studies) much easier.

DB files can be used by more than one job. For example, you may have the same exact demographic questions for each study. Once those question variables exist in a DB file, you can reference those definitions and use them as they are, or modify them (i.e., perhaps the column location has changed) and optionally save the changed variables in a new DB file. Up to ten user-defined DB files can be referenced at the same time during a single job.

The advantages of Mentor DB files are:

  • anything created by the program (variables, table elements, tables) can be saved into a DB file for future use;
  • items stored in a DB file can be retrieved quickly and easily;
  • tables stored in a DB file can be recalled and given minor modifications without having to reprocess them;
  • items stored in a DB file take up less disk storage space;
  • since complex elements are stored by a simple name, it is easier to generate ad hoc tables;
  • experienced users can generate large numbers of historical tables efficiently.

INTEGRATION OF SURVENT AND MENTOR

A principal feature of this software is the extensive integration of Survent and Mentor. For example, you can write a questionnaire to create your own user interface for Mentor. This means that it is possible to have a questionnaire, written for Survent, which asks a person what variables to tabulate, and then automatically produces the tabulation. (CfMC’s utility SCAN is an example of this.) Many other scenarios are possible, including retab systems, and specifications that interact with a specific database. You can design client-specific questionnaires with Survent, or CfMC can create questionnaires for your company on a contractual basis.

Because you can mix the modules of Survent and Mentor commands, complex, hierarchical studies are much easier to manage (this also called master-trailer processing). It is possible to write specifications in which there are master records (i.e., families), first-level detail information (i.e., family members), second-level detail information (i.e., trips made by family members), and so on. Each level of information has its own questionnaire, and during the interviewing process, Survent passes information to and from Mentor as necessary.

Any process that you need to use to capture data can be a Survent-Mentor application, including those not traditionally thought of as market research. Many companies can take advantage of the ability to do high-quality interviewing. For example, the software could be used with a list of clients, with each client having an array of tests to be done (some more than once). You could use CfMC software for the workers who need to review and manage the testing process. You can use Survent to control the content and location of informational text on the screen. Users see text written in language they understand, and you can reduce errors by controlling responses. Mentor manages the data. We call this ZTCS, Zero Training Custom Software, and it provides limitless possibilities.

EASE OF USE

CfMC software allows you to create readable, accurate, and fully labeled tables easily from a data file. Using the utilities, you can produce a wide range of tables just by responding to a set of menus. To indicate what information you want on your tables, you can use either data locations or names of variables from the study. You can also use the utilities for generating and manipulating data (including sorting, merging, and creating subsets from data files) and writing reports.

If you have Survent, it will take text and data locations from the original questionnaire and generate specifications that include all the basics of a table (the title, banner, stub labels) that you can use as is or modify. This way, you can produce a wide range of tables just by knowing the names of the questions to tabulate. For example, you can tell Mentor you want the answers from the question “OWN” cross-tabulated with the answers from the question “YESNO.” Here is the table Mentor would create:

 

TABLE 001

 

 

BANNER:4. Do you currently own or rent your home?

 

STUB:1. Do you, or any member of your household, have credit cards?

 

 

TOTAL   OWN    RENT    DK/NA

 

TOTAL 201 107 91 3

100% 100% 100% 100%

 

Yes 150 88 61 1

75% 82% 67% 33%

 

No 47 16 29 2

23% 15% 32% 67%

 

DK/NA   4 3 1 –

2% 3% 1%

 

Mentor is an extremely flexible tool. The menu-driven interfaces of the utilities allow its use by minimally trained staff, enabling the advanced spec writer time to develop more complex processes needed for specific applications.

1.3 THE BASIC STEPS OF THE REPORT/TABBING PROCESS

Processing a tabulation job typically consists of several key steps:

  1. Setting up the data file
  2. Defining the data descriptions
  3. Cleaning the data
  4. Generating new data from existing data
  5. Building and printing the report-ready tables

These steps are detailed in the following sections.

6. Setting up the Data File

Data files need to be in the CfMC System file format for Mentor to access it. Your data file can come from Survent or other interviewing or data entry software. Survent data files are already in the CfMC System file format. You can use the utilities to covert other types of files (ASCII, binary) to a System file.

When converting a data file, you can also sort it on very simple or complex criteria. You can also check for and make corrections for duplicate case IDs, duplicate data, or missing data. If you need to create a subset of the data (often used to test cleaning or table specifications), you can also select a specific portion of a larger data file (this is similar to using a base, but more efficient in terms of processing time).

7. Defining Data Descriptions

Data definition can be done in two different, but not exclusive, ways: use Survent’s DB and/or DEF files, or use command language to create specification files. The definitions consist of a variable name or label, question and response text, data location and width, and type of question. These elements are then used by the program when performing the other phases of the job.

Mentor is very efficient. Once data has been defined, the remaining phases – cleaning, generating, and table building – are able to use these data definitions. The DEF and DB files can be preserved, modified, and/or added to, allowing quick and easy processing.

8. Cleaning the Data

At the start of the cleaning process, you probably want to generate a holecount (marginal) to find out what is in your data file and get an idea about how much cleaning it requires. You can generate a holecount and other reports about the data with the utilities, see the Utilities manual.

Data cleaning involves systematically examining the data along with some logic specifications. With Mentor, you can clean your data interactively or in batch mode. (Data collected by Survent would typically need little, if any, cleaning.)

Specifications tell Mentor which data is valid and which have errors for each case. You can develop cleaning instructions dealing with linked responses (skip patterns, conditionals) or non-linked responses. Both will check for valid responses, check that no more than one of the responses was entered (for single-response questions), or might check to be certain that combinations of answers make sense. You can choose to have the program automatically “fix” the data for you all at once, or get a report of errors and fix them interactively.

Mentor’s cleaning commands are comprehensive. You can use IF-THEN-ELSE structures, and you can nest these structures. You can use GOTO statements for skip patterns or branching. You can also provide your own text for error messages to make error messages easier to understand.

9. Generating New Data

Once the data is clean, you may want to combine or otherwise manipulate your data to create new categories (this is commonly called recoding). Mentor provides a number of ways for combining data, performing arithmetic calculations and other data manipulation. You move data around to provide consistency between cases in different data files (i.e., across different waves of a tracking study). You can also combine data to form new items. For example, you might, for efficiency, want to create a single category for females over 35 who drive sports cars.

You can create new data by doing calculations on existing data items. For example, you may have numeric data on how many miles a car has been driven and gallons of gas it used. You can combine these figures to compute a miles-per-gallon figure. You can also add entirely new items to the data file, such as weighting factors or you can remove the rotation from data for series of questions that were rotated and stored in different fields.

10. Building and Printing Tables

Mentor produces tables using already defined question variables located in the DEF and DB files and, if necessary, more complex variables. These complex definitions can consist of any combination of other variables. You can use text-type variables to specify the format of items, such as titles, headings or complex banners. You can also use variables to dictate the printing options, such as whether to include vertical or horizontal percentages.

Variables are the fundamental units of processing with Mentor. The simplest variable consists of all of the valid answers to a single question from a questionnaire. You may already have variables from Survent. You can define your own or other variables, such as a base that includes only males, or a variable for specific numeric ranges.

Changing how the data appears on your tables is easy. Your options include whether or not to include frequencies, percents or statistics. Percentages can be based on a specific row, or you can have the percent base change mid-table. Percents can print with from zero to two decimal places of significance. Frequencies can print as whole numbers, or with one or two decimal places (for weighted data, for instance). You can also have your choice of different levels of decimal significance with generated statistics. You can have each row in the table print in a different format; for example, you can have some rows that have frequency and percents, but also have other rows that just print as frequencies (statistics, for example).

You can edit any text on your tables and reprint your tables without having to re- process the data in the tables. You can also use a variety of options to tailor the overall page format your tables. Again, these parameters can be changed between tables (i.e., do it one way for the first ten tables, do it another way for the next ten.). Options include the length of the page (number of lines per page), the width of columns in the banner, the width of row labels, the spacing between rows, whether or not to print summary rows and columns (Total, No Answer), and page numbering.

1.4 SUMMARY OF THE CHAPTERS

The chapters in this manual help you with the standard steps of a reporting or tabulation project. The exception to this is the supporting utility programs that are covered in a separate volume. We recommend that you read all chapters of this manual, even if you feel that you are already familiar with some of it or don’t need those particular features for your operation. Each chapter builds upon previous chapters, especially the table building chapters (Chapters 4, 5, and 6). Here is a short description of what is covered in each chapter:

Chapter One: Getting Started. This chapter provides a basic overview of Mentor and a summary of the reporting process.

Chapter Two: Cleaning the Data. Using Survent or your own variables, you can check and correct your data.

Chapter Three: Reformatting your Data. From your existing data, you can use the format of your data and create new categories of data for processing purposes.

Chapter Four: Basic Tables. This chapter teaches you how to produce simple tables. It also includes an explanation of the table defaults.

Chapter Five: Intermedite Tables. This chapter teaches you how to make changes to the default settings for tables, including printing settings.

Chapter Six: Complex Tables. This chapter explains how to add statistics (mean, standard deviation, and standard error) to your table, how to create top-box/bottom- box tables, how to weight your tables, how to use loop variables, and how to create break tables. It includes many examples of specification files and completed tables.

Chapter Seven: Customizing specifications for viewing tables via a browser. This chapter describes Mentor programs such as WebTables, On-Demand tables, etc.

Chapter Eight: Statistics. This chapter explains how to add T-tests, chi square, and ANOVA tests to your tables.

Chapter Nine: Specialized Functions. This chapter explains how to generate specialized reports and manipulate tables with System constants and special functions.

Appendices: The appendices are in a separate volume and cover several topics. Appendix A shows the formulas Mentor uses in statistical testing. Appendix B is the largest appendix, and it is a listing of all the tilde commands (tilde command begin with a “~”) including the syntax and a list of options for the command.

1.5 GETTING TECHNICAL SUPPORT

We hope you find this manual useful. CfMC is always looking for feedback about our software, our manuals, and our technical support. Below is a list of ways to contact us, depending on what your needs are.

 

Contact Information:

 

415-777-2922   Customer Support

 

http://www.cfmc.com

 

survox.support@enghouse.com   Customer Support